Azure/azure-sdk-for-java com.azure+azure-ai-projects_2.0.0

on GitHub

2.0.0 (2026-03-27)

Features Added

• Added getDefaultConnection(ConnectionType, boolean) to ConnectionsClient and ConnectionsAsyncClient for retrieving the default connection of a given type.
• Added connectionName parameter overloads to createDatasetWithFile and createDatasetWithFolder in DatasetsClient and DatasetsAsyncClient, allowing users to specify which Azure Storage Account connection to use for uploads.

Breaking Changes

• Methods across sub-clients were renamed to include the resource name for disambiguation (continuing the pattern from 2.0.0-beta.1):
  • DatasetsClient: listLatestVersion() → listLatestDatasetVersions(), listVersions() → listDatasetVersions(), deleteVersion() → deleteDatasetVersion(), createOrUpdateVersion() → createOrUpdateDatasetVersion()
  • IndexesClient: listLatest() → listLatestIndexVersions(), listVersions() → listIndexVersions(), getVersion() → getIndexVersion(), createOrUpdateVersion() → createOrUpdateIndexVersion(), deleteVersion() → deleteIndexVersion()
  • EvaluatorsClient: createVersion() → createEvaluatorVersion(), getVersion() → getEvaluatorVersion(), updateVersion() → updateEvaluatorVersion(), deleteVersion() → deleteEvaluatorVersion(), listVersions() → listEvaluatorVersions(), listLatestVersions() → listLatestEvaluatorVersions()
  • Same renames apply to the corresponding async clients.
• Connection.getCredentials() renamed to Connection.getCredential() (singular).
• ConnectionType.REMOTE_TOOL renamed to ConnectionType.REMOTE_TOOL_PREVIEW.
• EvaluatorMetric.setIsPrimary() renamed to EvaluatorMetric.setPrimary().
• BlobReferenceSasCredential.getType() now returns CredentialType instead of String.
• DatasetVersion.getDataUri() / setDataUri() renamed to getDataUrl() / setDataUrl() (also on FileDatasetVersion and FolderDatasetVersion).
• DatasetsClient.createDatasetWithFolder() no longer throws checked IOException; it now throws UncheckedIOException instead.

Bugs Fixed

• Fixed createDatasetWithFolder producing an invalid dataUri that caused a 400 error when registering the dataset.
• Fixed createDatasetWithFile using the dataset name as the blob name instead of the actual file name.

Other Changes

• Regenerated from updated API spec.